4 ms·
Algorithmically is it possible to pass only 'safe' interpreted code to the CPU? The majority of Javacript engines (Chakra, V8, Spidemonkey and JavasciptCore) ar
by petecox 9y ago
Algorithmically is it possible to pass only 'safe' interpreted code to the CPU? The majority of Javacript engines (Chakra, V8, Spidemonkey and JavasciptCore) are under a security-auditable FOSS license. Browsers claiming Spectre/Meltdown immunity would be in the next release notes, surely.
I'm not discounting that CPUs should operate correctly into the future but the current sandboxes would seem to rely on whims of a CPU's microcode. And the question for browser makers is whether to trust them.
- gruez 9y ago>Algorithmically is it possible to pass only 'safe' interpreted code to the CPU? in other words, the halting problem?
- amock 9y agoIt's not the halting problem because you can have false positives. If the code can't be proven safe it is rejected.
- kazagistar 9y agoIts not about algorithms. The issue is mentioned in the paper: even if the cache-based side channels are stopped (lets say, by manually clearing any cached memory as soon as its accessed, reducing performance to 0.1% of what it is now) there are other potential side channels like through limited CPU resources. A lot of the "fixes" might just be mitigations, that don't solve the problem entirely, but make it take much longer. Maybe you can cover up enough holes, and make the timers available poor enough that it takes a few hours or days to read a single byte. In other words, mitigations to make things less simple.